опыт трансформации старопромышленных городов / dis3778
.pdfto facilitate the deal decision-making process it needs to be specific, i.e. it must refer to a similar deal decision (Haleblian et al., 1999). As such, it requires decision makers to have been previously involved in a similar acquisition or divestiture decision. If decision makers do not properly discriminate their past deal experience, but generalize experience inappropriately this transfer might even harm the decision-making process (Haleblian et al., 1999). The risk for deal decision makers to inappropriately generalize their past experience can be considered high. Acquisition and divestiture decisions occur in many companies not as frequent as other strategic decisions (e.g. periodic capital expenditure, HR or marketing decisions). As a consequence, many decision makers have a rather moderate stock of past experience. In particular, these relatively inexperienced decision makers may be tempted to generalize from their small stock of experience too early (e.g. “an acquisition is an acquisition”).
Moreover, decision makers in deal settings have often been found to be subject to cognitive biases (e.g. Duhaime and Schwenk, 1985). In order to navigate through complex and ambiguous decisions, decision makers use simple analogies that lead them finally to an overly simple view of the situation (Steinbruner, 1974). Previous studies have shown that decision makers often make fatal mistakes when they leverage their experience based on superficial reasoning by analogy (Gavetti, Levinthal, and Rivkin, 2005). In view of the differences between the existing types and forms of deals, decision makers can quickly fall into this trap. For example, due its partial symmetry of activities managers often consider divestitures a “simple mirror image” of acquisitions. While this view certainly applies to some activities – as outline above – it applies not to the entire decision making process. For instance, due to the different perception of acquisition (“win”) and divestiture (“loss”) decisions, the requirements regarding the successful involvement, motivation and communication tend to be different between both decision types (Dranikoff et al., 2002). Decision makers who overlook these differences and base their experience transfer on simple analogies (e.g. “a deal is a deal”) are to encounter significant difficulties in the deal process. Vice versa, decision makers who rigorously discriminate their existing deal experience and apply their experience selectively to the present deal context can be expected to be more effective in their deal decision-making. Following this line of reasoning and previous empirical evidence on the positive performance effects of specific experience (e.g. (Haleblian et al., 1999), we expect decision makers’ specific deal experience to enhance the deal performance. Hence:
Hypothesis 2: In deal decision-making, decision makers’ past specific deal experience is positively related to deal performance.
15
Analysis-driven decision-making. In quadrant 3 of the matrix, a low level of specific deal experience among decision makers is followed by an intense analytical activity. If decision makers are confronted with an unknown complex strategic issue it is unlikely that the simple application of experience-based procedures will yield a satisficing result (Cyert et al., 1963). If past experience is not available or not relevant to today’s deal decision in a direct and straightforward way decision makers should expend resources to understand the decision context properly. In this case, decision makers may benefit from efforts to push investigatory activity to high levels since a stable core of norms and specific routines is missing. The creation of new knowledge and new insights consequently takes a critical role (e.g. Eisenhardt, 1989; Judge and Miller, 1991). Basically, it is this quadrant that accounts for the positive performance results of analytical intensity in the literature on strategic decision making.
The research on strategic decisions has suggested that the comprehensive search and analysis of information facilitates the proper understanding of situations and the making of good choices. While empirical evidence for transaction-specific settings is still missing, analytical intensity is to influence the outcome of portfolio decisions in a similar way for at least two reasons. First, analytical search helps to ensure that most or all important decision variables in the acquisition or divestiture setting are considered, resulting in better decisions and better decision outcomes. In acquisition or divestiture settings, decision makers have to face a number of difficult questions related to opportunities (e.g. value creation potential and attractiveness of potential targets or divestiture candidates) and risks (e.g. hidden “poison pills” in the portfolio of a target, underestimation of anergies in case of divestiture). A comprehensive information search and analysis is thus to help decision makers to evaluate these issues for each deal step in a systematic way. Take, for example, the identification of a potential acquisition (or divestiture) candidate. The analytical search and screening of the competitive (internal) environment can help decision makers to create long lists of potentially attractive candidates in line with corporate strategy and avoid an overly opportunity-driven approach with an exclusive focus on the firm’s immediate environment (the usual internal suspects) (e.g. Rankine, 1998). In a similar way, this approach may help decision makers in answering valuation issues on potential acquisition or divestiture candidates. While the explicit use of multiple analytic techniques, such as the DCF, multiple or net value asset method, is not to provide decision makers with a single, clear cut answer, it is to push them to make their assumptions explicit, think in different scenarios and also consider the probability of failure. Together, these exhaustive analytical activities ensure that all important input variables have been discussed and no relevant aspects have been overlooked thereby enhancing the decision quality.
16
Second, the analytical intensity helps decision makers reduce the risk of cognitive biases in their decision-making. Beyond the general difficulties inherent in managing a complex strategic situation, such as an acquisition or a divestiture, these biases often cause decision makers to search in the wrong places, overestimate wrong information, and ignore some important other effects (Duhaime et al., 1985). In consideration of the different types and forms of acquisitions and divestitures, a comprehensive approach could prevent decision makers from (mis-)applying analogies from apparently similar context and from finally adopting an overly simplistic view of the situation (Steinbruner, 1974). Moreover, it could help decision makers avoid an illusion of control since decision makers are often tempted to overestimate the extent to which the outcome of a deal is under their personal control. For example, in a recent study Malmendier and Tate (2005) illustrate how overconfidence can lead decision makers to overestimate the returns of their investment projects and view external funds as unduly costly thus heavily distorting the corporate investment activity. Moreover, comprehensiveness may avoid decision makers to engage in single outcome calculation (Duhaime et al., 1985). For example, in the divestiture context it can enable decision makers to specify all alternative courses of action for dealing with a failing unit instead of focusing too early on a single goal and a single alternative (divestiture) for achieving it (Duhaime et al., 1985). With such biases having substantial effects on the outcome of acquisition and divestiture decisions, analytical intensity likely improves the chances of reasonable information gathering and processing in a deal context if no deal experience exists.
Building on this reasoning, we expect analytical intensity to facilitate the proper understanding of ambiguous and complex deal contexts and to minimize the risk of cognitive biases in the deal decision-making. This more effective deal decision-making process is to lead to superior deal outcomes.
Hypothesis 3: In deal decision-making, analytical intensity is positively related to deal performance.
Over-investigated decision-making. In quadrant 2, decision makers engage in high investigatory activity even though deal experience exists. From behavioral theory we know that decision makers have “limited information, attention, and processing ability” (Greve, 2003: 12). Faced with a complex strategic issue, such as a portfolio deal, decision makers are per se typically “confronted with more stimuli [they] can attend to or adequately process” (Hambrick, Finkelstein, and Mooney, 2005: 478). As a result, if decision makers are unable to satisfice – that is to look for a course of action that is satisfactory rather than optimal (Simon, 1945) – by drawing on available experience that economizes on information
17
processing, they will quickly come to their cognitive limits (Bromiley, 2005). While decision makers might be tempted to complement insights from past experience through intense investigatory activity in order to gain confidence, research in cognitive psychology suggests that this approach is to trigger significant problems. This effect may materialize in the presence of both specific deal experience and general deal experience.
Specific deal experience. Decision makers who have been already involved in similar deal settings in the past are to profit from their past experience. Experienced decision makers are to have a sound grasp of aspects regarding the procedure as well as the content of the deal decision. As such, they possess the information and expertise required to address the critical aspects of acquisition and divestiture opportunities. If the decision makers engage – despite their specific deal experience - in heavy investigatory activity to complement their insights, they can be expected to produce not only new, but also redundant information. This redundant information would not hurt as long as it leaves the decision-making behavior unaffected. Previous research in cognitive psychology suggests, however, that the use of redundant information in decision-making is to deteriorate the quality of the decision outcome (Davis et al., 1994; Kahneman et al., 1973; Nisbett et al., 1981). In their study on individuals’ prediction about target individuals, Nisbett and colleagues (1981) found that the predictions of subjects provided with a mix of diagnostic and non-diagnostic information were much less accurate than those of the subjects holding only diagnostic information. The non-diagnostic information “diluted” the implications of diagnostic information (e.g. Kahneman et al., 1973; Nisbett et al., 1981). This dilution effect suggests that decision makers have significant difficulty in dealing with redundant, useless information when mixed with relevant information. Moreover, Davies and colleagues (1994) suggest that it is not only redundant, but also abundant information that negatively affects decision makers’ effectiveness. In their study on the forecasting of stock earnings, they found that information provided in addition to common baseline information is to decrease decision quality independent of its redundancy. At the same time the use of redundant information increased the confidence of decision makers in their judgments.
In the challenging context of acquisition and divestiture decisions, these effects may even reinforce themselves and lead deal decision makers to get trapped in a “vicious” circle of increasing confidence, but decreasing decision quality. In face of the significant ambiguity, complexity and lack of structure deal decisions entail, decision makers may quickly feel uncertain about the sufficiency and appropriateness of existing (experience-based) insights (“Do we have all information we need on the target?”, “Have we really understood all relevant facts of the industry”?). In order to lower their uncertainty and strengthen their confidence decision makers may start to reach out for additional and ideally new information. While this is to enhance decision makers’ confidence in their judgments it may likewise
18
negatively affect decision quality (Davis et al., 1994; Nisbett et al., 1981). The decisions may thus not yield acceptable results from the point of view of evoked goals and aspirations (Cyert et al., 1963). This unsatisfactory feedback is likely to lead decision makers to intensify their search and analysis activities and to exacerbate the evoked effects.
Building on this line of reasoning, we expect in settings in which decision makers possess specific deal experience, heavy investigatory activity not to strengthen, but to weaken the quality of the decision-making process and finally the deal performance. Hence:
Hypothesis 4: In deal decision-making, the combination of past specific experience with intense analytical intensity is negatively related to deal performance.
General deal experience. In a similar vein, the existence of past general deal experience is to increase the risk for decision makers to produce with their investigatory activity redundant or abundant information. As decision-making in a deal context is repeated over time, the main activities in the process become increasingly routinized (Levitt et al., 1988). These routines get embedded in the organization and are to help decision makers address the challenges they encounter in the deal decision-making processes. If in this context decision makers reach out for additional information and engage in heavy investigatory activity this may not only lead to valuable and complementary insights, but is also to produce redundant and unnecessary information. As outline above, we suggest this production of potentially redundant information through heavy investigatory activity to decrease the quality of the decision outcome (Davis et al., 1994; Kahneman et al., 1973; Nisbett et al., 1981). The potentially redundant and abundant information risks to weaken the insights derived based previous deal experience thereby harming the deal outcome (e.g. Kahneman et al., 1973; Nisbett et al., 1981). As a consequence, we expect in settings where organizations possess previous deal experience heavy investigatory activity not to enhance, but to deteriorate the quality of the decision-making process and finally the deal performance. Hence:
Hypothesis 5: In deal decision-making, the combination of general deal experience with intense analytical intensity is negatively related to deal performance.
Under-investigated decision-making. In quadrant 4 of the matrix, decision makers suppress investigatory activity even though no specific deal experience exists. Since decision makers are confronted with an unknown complex strategic setting, existing routines and problemsolving procedures are to provide only little guidance on the effectiveness of potential courses of action. If decision makers nevertheless focus their efforts to identify satisfactory courses of
19
action based on solutions adopted in the past, the decision outcome cannot be expected to yield acceptable results (Cyert et al., 1963). For example, if decision makers have no prior acquisition experience, but simply apply their routines and procedures developed in the management of internal investment decisions to an acquisition context, the acquisition outcomes may fall behind expectations. Competitive bids, for instance, often require decision makers to cope with missing or incomplete information under severe time constraints. Not to calibrate the decision-making process accordingly, but to follow a careful and lengthy reexamination procedure similar to internal investment decisions would close valuable windows of opportunity and leave management with a potentially sub-optimal set of investment options.
Hence, if decision makers can neither draw on specific experience nor engage in a more comprehensive, analysis-driven decision-making style the outcome of their deal decisionmaking is simply subject to luck (Barney, 1986). As previous research suggests this circumstance might not per se result in inferior performance outcomes (Barney, 1986). However the odds are working against it – in particular in deal settings. Numerous studies and reports have highlighted the high risk and failure rate associated with deal decisions (e.g. King et al., 2004). Consequently, if inexperienced decision makers make no effort to understand these complex settings through comprehensive search for opportunities, formal codification of generated options, and intensive analysis, the probability is high that their decision outcome will yield unsatisfactory results.
From the reasoning for hypotheses 1-3 it follows indirectly that inexperienced and uninformed decision makers will yield deal outcomes inferior to those of their experienced and informed peers. Notwithstanding the fact that also inexperienced and uniformed decision makers might be lucky and realize favourable deal outcomes, overall the odds are small and the complexity of the portfolio restructuring decision is working against a superior performance outcome based on good fortune. Since already included in the first three hypotheses, we refrain from stating this effect in a formal hypothesis.
2.4 METHODOLOGY
For the purpose of this study, we take the single portfolio restructuring decision as unit of analysis. This is in sharp difference to most previous decision-making studies which conducted their analyses at the firm-level (e.g. Fredrickson, 1984). Firm-level analyses, however, are problematic for several reasons. First, when operating at the firm level the direct linkage of decision processes to firm performance is problematic. The causal ordering is ambiguous and the relationship between decision process characteristics and firm performance is likely to be confounded given the many endogenous and exogenous effects on organizational performance (Pearce, Freeman, and Robinson, 1987). Moreover, firm-level
20
analyses assume that decision processes are uniform across decision types as well as across an entire organization (e.g. Glick et al., 1993; Goll and Rasheed, 1997). But as earlier studies have shown (Elbanna et al., 2007a; Elbanna and Child, 2007b; Hickson et al., 1986; Papadakis, Lioukas, and Chambers, 1998) even within the same organization decision processes are very likely to differ from one decision to another. Taking the individual strategic decision as unit of analysis thus enables a closer match with reality and ensures a much more robust analysis of the influence of decision-making process characteristics and (financial) decision outcome.
Data Collection
The choice of sample for our survey was dictated by the need to identify companies that have undertaken acquisition or divestiture decisions within the last two years. We started with the population of 110 publicly listed companies on the three major German and Swiss stock indices: Deutsche Aktien Index (DAX 30), the Midcap DAX (MDAX 50), and Swiss Leadership Index (SLI 30) as of February 2008. We then identified those companies that had made acquisitions and/or divestitures during the previous two years (2006-07) as provided in Thomson Onebanker’s transaction database. This resulted in an initial sample of 102 companies, listed on one of the three major German and Swiss exchanges, with acquisition and/or divestiture activity during the 2006-07 time period.
We contacted 2 an executive (Chief Financial Officer, Head of M&A, Head of Corporate Development or Head of Corporate Strategy) at each of these companies by email and phone to solicit their participation in our survey. Of the 102 companies, 60 (59%) agreed to participate in the study. Of the 60 companies that agreed to take part in our survey, we received responses from 56 companies, resulting in a response rate of 93% and an overall participation rate of 55% (based on our initial sample of 102 companies with acquisition/divestiture activity). Compared to studies on mergers and acquisitions (Capron and Shen, 2007) and on decision-making processes (Atuahene-Gima et al., 2004), this level of survey participation is very high. The four companies that had initially agreed to participate in our survey, but did not respond, were involved in an acquisition at the time of the survey. The 56 participating companies represent 20 different industries, with no industry accounting for more than 20% of the sample. The most predominant industries were industrial goods (20%), financial services (16%) and pharmaceutical and chemical goods (14%). The companies ranged in size from 1,150 to 525,000 employees, with an average of 62,300
2 For our data collection, we collaborated with a leading management consultancy. Both our university and the consultancy had established contacts at each of these companies through their previous participation in the executive education and consulting services respectively. To identify and contact the executives at the different firms we further relied on the extensive alumni network of our university and the consultancy. Alumni at the respective firms were able to provide us with the contact information of the executives. We refer to these executives as our “executive contact”.
21
employees and from $0.85 billion to $138 billion in sales, with an average of $26.6 billion in sales. To investigate whether our study was subject to non-response bias, we compared our sample of 56 companies with the 46 companies that did not participate in our survey by testing for differences in total sales (t = -0.33; p = 0.81), number of employees (t = 1.2; p = 0.29), and return on assets (t = 0.37; p = 0.86) and found that our sample is not statistically different from the 46 companies that chose not to participate in the study. We also tested for late respondent bias and did not find any statistical differences between early and late respondents within the sample.
Since our study’s unit of analysis is a single acquisition or divestiture decision we asked our executive contact at each of the companies to name one or two acquisition or divestiture decisions taken within the previous two years. For the 56 companies in our sample, 80 acquisition and divestiture decisions were identified. We chose to utilize two respondents for each of the 80 acquisition or divestiture decisions, because reliance on a single respondent when acquiring organization data is prone to various forms of biases (Cote and Buckley, 1987; Podsakoff et al., 2003; Podsakoff and Organ, 1986; Rindfleisch et al., 2008; Van Bruggen, Lilien, and Kacker, 2002) and the use of multiple respondents has been shown to produce data of superior quality and validity (Van Bruggen et al., 2002; Zhou, Shin, and Cannella, 2008). After identifying the specific strategic decision, we asked our executive contact at each of the companies to name two managers who had intimate knowledge of the decision process through their involvement in the acquisition or divestiture decision. We notified each of these managers that they had been selected by our executive contact to participate in our survey and forwarded them the survey instrument via email.
Since strategic awareness of acquisitions and divestitures has been shown to be positively related to hierarchical level in most companies (Hambrick, 1981), we limited our set of survey respondents to top management team members, members of the corporate center, and business unit top management since these managers would be the most aware of strategic decisions of this nature. We verified their executive position as part of the survey.
In total, we collected 160 completed survey questionnaires – two respondents per firm for each of the 80 acquisition and divestiture decisions representing 56 companies.
Operationalization of variables
Dependent variable
Following a growing stream of work (e.g., Bruton et al., 1994; Hayward, 2002; Kale, Dyer, and Singh, 2002; Kale and Singh, 2007) and partly in response to increasing criticism of abnormal stock returns (CAR) as a performance measure (e.g., Oler, Harrison, and Allen, 2008; Zollo and Meier, 2008), we decided to use a perceptual measure as our dependent variable. As the primary interest in acquisition and divestiture research is on financial
22
performance we designed our performance measure accordingly. We asked managers to assess the extent to which the acquisition/divestiture met top-line (i.e. revenues), bottom-line (i.e. budget) and productivity (i.e. efficiency) objectives (for a detailed description of the items, see the Appendix). All items were measured on seven-point Likert scales. The Cronbach alpha for our measure was .85 in the overall sample indicating strong reliability.
Independent variables
General deal experience. We conceptualize a firm’s general deal experience as the amount of knowledge or skills a firm has developed through its past deals completed. Building on previous experience studies (e.g. Haleblian et al., 1999; Haleblian et al., 2006; Hayward, 2002; Ingram and Baum, 1997), we measure a firm’s deal experience by the number of deals the sample firms completed during the past two years. As such, we do not distinguish between the different types of deals, but capture the firm’s deal experience overall. We obtained our data on the firm’s past deal statistics from the Thomson Onebanker database.
Specific deal experience. We conceptualize specific deal experience as the amount of deal-specific knowledge or skills the group of decision makers has acquired through its participation in similar deal decision-making processes. We thus measure the amount of specific deal experience based on decision makers’ previous exposure to similar deal settings. For this purpose, we asked each respondent to report the intensity of his or her previous involvement in (a) similar deal decision(s) (for a detailed description of the item, see the Appendix). We measured this item on a seven-point Likert scale.
Analytical intensity. There are several aspects that make for analytical intensity in decision-making processes. First, as Sharfman (1997) points out, “the extent to which the decision process involves the systematic collection, analysis, and use of information” underlies more analytical decision making (1997: 181-82). Moreover, analytical intensity has also been associated with the development of alternative courses of action and higher levels of systematic analysis using multiple criteria and tools in order to screen and evaluate alternatives (Fredrickson, 1984; Miller, 2008; Miller, 1987). Thus analytically intense decision-making processes involve both the collection and use of information and analytical techniques as well as the evaluation of alternative courses of action (Miller, Burke, and Glick, 1998). To fully capture the construct of analytical intensity, we created a measure composed of seven items drawn from the prior literature (Dean et al., 1996; Fredrickson, 1984; Miller et al., 1998; Papadakis et al., 1998; Talaulicar, Grundei, and von Werder, 2005).
All seven survey items utilized in our measure of analytical intensity were measured on a 7-point Likert scale. The Cronbach alpha for our measure of analytical intensity utilizing these seven items was .72.
23
Controls
Relative deal size. Deal size has been found to influence both acquisition and divestiture performance (e.g. Asquith, Bruner, and Mullins, 1983; Gupta and Misra, 2007; Moeller, Schlingemann, and Stulz, 2005). Moreover, deal size could also affect analytical intensity, thereby creating a spurious correlation between analytical intensity and performance. Relative deal size is measured using an ordinal variable: a “1” denotes a deal representing less than 5% of total firm sales; a “2” represents a deal that is greater than 5%, but less than 10%, of firm sales; a “3” represents a deal that is greater than 10%, but less than 20%, of firm sales; and a “4” represents a deal that is greater than 20% of firm sales.
Deal importance. Previous decision-making research has shown that the perceived magnitude of impact of a strategic decision is among the strongest explanations of decisionmaking behavior and outcome (e.g. Elbanna et al., 2007a, 2007b; Papadakis et al., 1998). In order to assess the deal’s importance respondents were asked to evaluate the perceived importance and visibility of the decision within their firms. The two items were measured on a seven-point Likert scale. The Cronbach alpha for our overall sample was .78.
Deal relatedness. Previous transaction research has shown, e.g. in acquisition contexts, that business relatedness of the target firm with the focal firm can have a significant impact on the deal outcome (e.g. Capron et al., 2007). In order to assess the relatedness of the deal respondents were asked “Was the strategic decision related to the company’s primary business?” Respondents could answer “yes” or “no”.
Firm performance. Acquiring and divesting firm performance appears to be positively related to deal success (Morck, Shleifer, and Vishney, 1990). We measured firm performance by subtracting the median ROA value from the firm level value for the relevant transaction period. We obtained this data from the Thomson Onebanker database.
Firm size. Many researchers have argued that firm size can affect analytical intensity (e.g. Fredrickson et al., 1989; Miller et al., 1998; Simons et al., 1999; Snyman, 2003), such that larger firms will employ more formal and analytic processes (e.g. Papadakis et al., 1998). We use the (logarithmic) net sales of each firm in the year prior to the focal acquisition/divestiture decision to control for firm size. Sales data was obtained from Thomson Onebanker.
Validity and Reliability
We took several measures to assure validity and reliability in the survey construction and its administration. To ensure validity of the data we conducted a pre-test of the survey questionnaire using three partners and four experienced merger and acquisition project managers of leading management consulting companies. We met individually with these experts to get feedback on questionnaire construction and to refine the wording of the survey questions. Specifically, we asked these management consultants to utilize the most recent
24